Anyone know why the predictive routing engine seems to ignore the new schedule adherence data right after a weekly publish? we are in ct (america/chicago) and our schedules go live every monday at 12am. for the first 2 hours, the queue depth calculations for our inbound sales queue are way off. agents are marked as available in wfm but the predictive router thinks they are on break or unavailable. this causes a huge spike in abandon rates because the system underestimates the workforce capacity.
i checked the /api/v2/wfm/schedules/{scheduleId} endpoint and the status is published. the agent states in the admin console show them as logged in and ready. but the predictive routing dashboard shows a utilization rate of 0% for those agents until about 2am. we are using the default predictive routing strategy with no custom skills assigned to this queue. is there a known sync delay between the wfm schedule publish and the predictive routing workforce model? or did we miss a configuration step in the architect flow for handling schedule updates? this is killing our service level compliance on mondays.